Programming AVRs without AVR Studio?

[paulie]

yes, all 3 exe files including avrisp(sp?) worked outside the environment. i have no idea where they are located on your disk. as i said it took me hours to find it on my own hard drive. why are you avoiding the obvious solution, avrdude.exe, which is used by 99.9% of avr programmers and is available from literally hundreds of internet sites?

[Erwin Ried]

khazama avr programmer kicks avrdudess buttocks
http://khazama.com/project/programmer/

[paulie]

its shouldnt be hard to kick avrdude butt because there are many things still need fixing and no interest in the creator to do so. ie those totally wacko fuse listings. and many errors in the error messages. however unless kazama has command line capability for reasons of productivity and scripting then im afraid there is no competition for serious programming. and what about the device list? will it do tpi? >128k flash?

[dannyf]

-it  shouldnt be hard to kick avrdude butt because there are many-

the other poster was talking about kicking avrdudess butt, not avrdude butt.

avrdude is very different from avrdudess, like apple. Vs. Orange.

[paulie]

ah... i didnt read closely. so kazama is just another pretty dress. then maybe more like apple and apple skin.

anyway im still waiting for something better than dude to come along or fix the big problems. maybe theres hope... finally after years of waiting latest release does skip ff locations instead of pointlessly wasting several minutes programming a tiny 512 byte bootloader up high.

[dannyf]

A few other solutions:

1) bootloader: FLIP from Atmel is feasible but not as good. Teensy's bootloader is excellent but only available on Teensy. You have the Arduino stuff too. I think Atmel publishes a few of its own. But I don't think they work on 100% of their chips. And they are not practical on smaller chips.

2) Dragon: It has the added benefits of hv programming.

3) pickit2: it can be used to program avrs (just some?). Not sure of the limitations.

4) write your own: it gets the job done exactly the way you wanted.

...


To me, programming the chip is so trivial that I care little about the programmer / programming software. I guess that's why you see such a lack of interest / progress / bug fixes.

[paulie]

3) pickit2: it can be used to program avrs (just some?). Not sure of the limitations.

part of my job as firmware engineer is developing bootloaders. many different types for at least 3 of my clients so im pretty up to speed on what is and is not available out there. so flashing tools are far from trivial for me. little things like that ff issue have a HUGE impact on my productivity to the tune of hundreds of man hours wasted (no exaggeration). so important i was forced to fix the problem myself but so many other bugs in the version i started from it was necessary to maintain two complete sets of exe and conf. one for bootloader work and another that works properly with all the avrs. until recent release that is.

as far as pickit2 limitations, cost tops the list. you can literally buy a dozen avrasp which do ALL atmel chips instead of just a few like microchip. very important if you need to supply many clients and entire classrooms. thankfully pic is fading from both the hobby and academic scene.

[zapta]

I tried to program a new atmega328p chip with avrdude + avrdudes but it did not work. It could not even read the signature (got all zeros).

I looked at the ISP signals and the clock sclk was about 1Mhz. I read somewhere that virgin AVR  chips come factory fused to 1Mhz internal RC clock and that the ISP clock speed can be at most 1/4 of that frequency (that is, no more than 250Khz).

Anybody know how to tell avrdudes to reduce the ISP speed?  Avrdude has a '-B 4' option but I don't know how to make avrdedes to use it.

I have no problem programming chips that already oscillates at 16Mhz.


Edit: figured it out. AVRDUDESS has a Bit clock (-B) text field in the main menu. Setting it to 4 does the trick. It's kind of strange that the default setting is incompatible with out of the factory blank chips.

[dannyf]

flashing tools are far from trivial for me...

This is where I have trouble understanding you. You apparently have the demand for a better programmer; You said you also have the skills to make one yourself.

Yet, you insist on using an inferior tool that doesn't fit your needs / stature.

Why not just make your own "avrdude" the way you want? It helps your productivity; and it helps the community if others may find your "avrdude" more desirable.

Instead, you are wasting your time complaining about someone else not having sufficient interests and urgency to fix a tool for you.

Where is the disconnect?

[Kremmen]

Sheesh. You are talking about hundreds of wasted hours messing around with flaky hobby tools. Man, the business case is obvious to anyone who can put 2 and 2 together. So run, not walk, to the nearest browser and order a proper AVR tool to handle the job for you. ISP Mk2 or 3 comes to mind and case closed. Or are all thos hours free of charge so that there is no matter how many you waste...?

[paulie]

Sheesh. run, not walk, to the nearest browser and order a proper AVR tool to handle the job for you. ISP Mk2 or 3

Why not just make your own "avrdude" the way you want? It helps your productivity; and it helps the community if others may find your "avrdude" more desirable.

there does seem to be a problem understanding, maybe language barrier, so i will repeat. i do have my own "dude" and it functions perfectly for all my professional needs. only on the rare occasion when forced to work in windows or other gui os do the real avrdude problems crop up and since that is not often and im aware of the bugs, no problem.

and im not looking for another programmer. usbasp is perfect. i have mkII, dragon, a $2000 dataio, and others sitting on the shelf and they dont hold a candle up to it. nothing but problems. as far as mkII i dont know how many times driver and firmware updates were required and probably still dont support tpi. reminds me of microchips programmers. not to mention you can buy a dozen usbasp for the cost of one mkII. so again, no problem.

i was merely commenting on some issues gui fanboys might encounter. even if i did publish my utility unlikely anyone would be able to use it. for most real mode cmd line os is like garlic to a vampire. this thread made that clear. anyway its good you are happy with status quo.

[Kremmen]

Oh, sorry, i'm sure. Whatever floats your boat.

[redben]

I pesonnaly use X Code on my macintosh with avrdude to get hex file.
The I upload hex with a very nice gui called avr tools : http://www.w7ay.net/site/Applications/AVR%20Tools/index.html

This app make the fuse set up much more user friendly.
Regards,

[dannyf]

only on the rare occasion

not often

no problem.

so again, no problem.

So you spent pages and days rigorously complaining about an issue that a) is actually a non-issue and b)  you rarely encounter?

What happened to the "HUGE impact" on your productivity, or those "hundreds of man hours wasted (no exaggeration)"?

Which side of the story are we supposed to believe?

usbasp is perfect.

Absolutely agree, as long as you exclude issues that it cannot solve.

[paulie]

lol. you are easily confused. i can go into a detailed timeline for last 7 years but lets just say i started out with the same tools you and most others here adore then, finding them inadequate, took matters into my own hands. no regrets because it did nudge me into a higher productivity environment. and only recently did the dude dude wake up and fix that ff bug which was the biggest time sink. now if he would only straighten out the mislabeled fuse messages and also stop calling 1 bits 0 it would be quite a nice tool.

issues that usbasp cannot solve? how can you solve "issues" that dont exist. or maybe you refer to non-hardware related.

[dannyf]

you are easily confused.

Sure, particularly when your statements are diametrically opposite of each other. While they may be both wrong, they cannot be both correct.

All I was trying to understand is how could you have trivial issues that cost you greatly in productivity losses and hundreds of man hours wasted? There can be only one possible explanation here and it doesn't bode well for you.

[dannyf]

issues that usbasp cannot solve? how can you solve "issues" that dont exist.

They may not exist to you. That doesn't mean that they don't exist at all.

All it takes is to know it better.

[paulie]

I was trying to understand is how could you have trivial issues that cost you greatly in productivity losses and hundreds of man hours wasted? There can be only one possible explanation here and it doesn't bode well for you.

you are making things up danny. i dont think i every said "trivial" issues impacted my productivity.

1. one particular NON-trivial problem cost me, by estimation, nearly a thousand hours of wasted time watching avrdude progress bars.  mostly loading programs that totaled less than 1/2 kbyte (500 bytes).

2. solved by developing tools that did not suffer this NON-trivial problem. also relieved my workload bailing out students and fellow club members.

3. years go by and avrdude developers start listening to myriad complaints (not me) and made changes. issues still exist but most of those who are familiar with the tool have workarounds so might be considered cosmetic or trivial. still causes noobs to brick chips so maybe not trivial to some.

its hard put it more simply but if you really want to continue this "discussion" i will try.

[dannyf]

i dont think i every said "trivial" issues impacted my productivity.

You did say that those issues were "no problem" 3x, and that they only impacted you on rare occasions.

That sounds very trivial to me.

[zapta]

Guysl, this is a useful thread, please don't get it locked.

[paulie]

agreed. silly that a person cannot understand there may be both nontrivial and trivial issues and the latter have little effect on productivity. maybe trolling, maybe personality issues, but i think out of hundreds browsing this thread it was worth repeating if the light bulb went on for a few.